Mistakes

Costly Mistakes in Wire Harness and Cable Assembly Estimating and Build

The wire harness and cable assembly errors that quietly blow up quotes and yields, from cut-length service loops to under-counted crimps, each with a symptom, root cause, and a numeric fix.

Symptom: your quoted wire footage runs 8 to 15 percent short and every build eats into buffer stock. Root cause is a cut length taken from the drawing dimension only, ignoring service loops, strip length, and routing slack. A harness with a straight-line branch of 640 mm needs roughly 25 to 40 mm of strip per end plus a 5 to 10 percent slack allowance on curved routes, so the real cut is closer to 700 mm. Fix: drive every cut from a Wire Cut Length calculation that adds strip, loop, and slack as explicit fields, then reconcile the summed footage against your spool draw-down each week.

Symptom: labor estimates look tight on paper but the floor misses standard by 20 percent. Root cause is under-counting terminations. Estimators price one crimp per wire when a typical wire has two ends, and splices, ferrules, and heat-shrink each add cycles. A 60-wire harness is not 60 crimps, it is 120 plus 8 to 12 splice operations. At 12 to 18 seconds per hand crimp or 4 to 6 seconds automated, that miscount is 15 to 25 minutes per unit. Fix: load a Crimp Labor Load count from the actual termination list, not the wire count.

Symptom: scrap cost shows near zero in the quote but the P&L says 4 to 7 percent of material is gone. Root cause is treating scrap as a flat 2 percent adder instead of modeling setup waste, mis-strips, and failed crimps that get cut off and re-terminated. Each pull-to-length setup on a cut machine wastes 150 to 300 mm, and a 3 percent crimp reject rate on a 120-crimp harness means roughly 4 rework cuts. Fix: run a Cable Scrap Cost estimate per SKU and separate one-time setup scrap from recurring per-unit scrap.

Symptom: test throughput collapses on high pin-count assemblies and the line backs up. Root cause is estimating continuity test time as a fixed number per harness rather than scaling with net count and connector cavities. A 40-net board tests in well under a minute, but a 400-net harness with guided probing and hi-pot can run 4 to 9 minutes including fixture load and unload. Fix: size the workload with a Continuity Test Workload calculation keyed to net count, then add fixturing time separately in Harness Test Time so labor and takt line up.

Symptom: connector and terminal spend overshoots the BOM by 10 to 20 percent. Root cause is pricing the catalog part number and forgetting the multipliers: seals, cavity plugs, secondary locks, and terminals priced per cavity not per connector. A 24-cavity connector may list at 3.20 dollars but carry 24 terminals at 0.11 to 0.35 each plus 6 unused-cavity plugs, pushing the loaded connector past 9 dollars. Fix: build Connector Cost and Terminal Cost line items at the cavity level and include seals and plugs as their own quantities.

Symptom: two shops quote the same print and land 30 percent apart. Root cause is inconsistent routing and board complexity assumptions. Routing labor on a flat single-trunk harness might be 4 to 6 minutes, while the same wire count on a 3D formboard with 5 breakouts and tape-and-tube coverage runs 12 to 20 minutes. Estimators who skip the Routing Labor step bury this in a vague overhead percentage. Fix: quantify breakout count, covering type, and formboard complexity explicitly so the routing minutes are defensible rather than guessed.

Symptom: unit conversions silently corrupt the whole estimate. Root cause is mixing AWG with mm2, inches with mm, and per-1000-foot spool pricing with per-meter cut lengths. A 0.35 dollar per foot wire is 1.148 dollars per meter, and a 5 percent slip here compounds across 60 wires. Cross-check by confirming AWG-to-mm2 mapping, for example 20 AWG is 0.5 mm2 and 16 AWG is 1.3 mm2, before any cost rolls up. Fix: lock one unit system per project and validate spool price basis against cut basis on the first article.

Symptom: the first production run yields 85 to 90 percent when the quote assumed 98 percent. Root cause is estimating from ideal cycle times with no allowance for learning curve, operator variation, or documented rework. Early builds run 20 to 40 percent slower than steady state, and a harness with 120 crimps at even a 1 percent per-crimp defect rate averages more than one rework touch per unit. Fix: hold a labor contingency of 10 to 15 percent for the first 100 units and reconcile actuals against the Cable Assembly Labor estimate before locking a repeat price.

Published 2026-07-01.